44S Proceedings of the NASS 28th Annual Meeting / The Spine Journal 13 (2013) 1S–168S

Chi-squared statistic, Fisher’s exact, and single sample t tests were used to

examine the data. Clearance of the infection was defined as normalizing of

serum markers and resolution of osteomyelitis on MRI after 6 months of

treatment.

RESULTS: One-hundred six patients meet the inclusion criteria specifi-

cally for the management of spinal osteomyelitis: 62 men (58%), 44 women

(42%), mean age 54 yrs., mean follow-up 38 months. Sixty-four patients

(60%) had paravertebral collections and 33 patients (31%) had epidural col-

lections. In regard to resorption of the collection, higher clearance with epi-

dural collections were observed compared to paravertebral ones [OR: 1.5,

95% CI: 0.6 to 3.6; p50.3]. Long term improvement in Oswestry scores

was less in the epidural group (from 61.5 to 50) compared to the paraverte-

bral group (from 66.3 to 44.5). The epidural collection group had a higher

ESR on presentation [63.6 vs 11.8]. Mean volume of the collection on MRI:

epidural 115 cm3, paravertebral 156 cm3, p50.5. The epidural collections

tended to enhance with contrast more often than paravertebral collection

[OR: 2, p50.11]. The epidural collections less than 100 cm3 tended to re-

sorbed more often and trended toward significance [OR: 3, p50.08].

CONCLUSIONS: The epidural collections were smaller and tended to en-

hance with contrast, therefore they resorbed more often than paravertebral

collections. For fluid collections that enhance with contrast or ones that are

smaller than 100 cm3, consider antibiotics alone. In vertebral osteomyelitis

with an ESRO 55, consider the presence of an epidural collection. Lastly,

epidural collections have less long-term improvement in Oswestry scores

compared to paravertebral collections.

FDA DEVICE/DRUG STATUS: This abstract does not discuss or include

any applicable devices or drugs.

http://dx.doi.org/10.1016/j.spinee.2013.07.135

88. Independent Predictors of Failure of Nonoperative Management

of Spinal Epidural Abscesses

Sang D. Kim, MD, MS1, Rojeh Melikian, MD2, Kevin L. Ju, MD3,

David Zurakowski, PhD4, Kirkham B. Wood, MD5, Christopher M. Bono,

MD6, Mitchel Harris, MD, FACS3; 1St. Louis, MO, US; 2Cambridge, MA,

US; 3Brigham & Women’s Hospital, Boston, MA, US; 4Boston Children’s

Hospital, Boston, MA, US; 5Massachusetts General Hospital, Boston, MA,

US; 6Brigham & Women’s Hospital, Department of Orthopedic Surgery,

Boston, MA, US

BACKGROUND CONTEXT: Spinal epidural abscesses (SEA) have been

traditionally treated with urgent surgical decompression followed by long-

term intravenous antibiotics. The notion that all patients with SEA require

surgical decompression has been recently challenged by reports of success-

ful medical management of select patients with SEA. To avoid subjecting

patients to potentially inadequate treatment with detrimental outcomes, the

independent variables that stratify the risk for failure of medical manage-

ment of SEA must be identified.

PURPOSE: The aim of this study was to identify the independent

variables that determine success or failure of medical management for

SEA.

STUDY DESIGN/SETTING: All patients admitted to the authors’ health

care system with a diagnosis of SEA from 1993 to 2011 were identified

and the data were retrospectively reviewed.

PATIENT SAMPLE: Patients 18 years of age or older diagnosed with

SEA documented by MRI or CT myelogram with minimum 2 months

follow-up were included. Excluded were those with postsurgical spinal

epidural abscesses or spondylodiscitis/osteomyelitis with an associated

phlegmon and those with a complete spinal cord injury from SEAwho pre-

sented forty-eight hours after the onset of paralysis.

OUTCOME MEASURES: Successful medical treatment was defined as

eradication of the infection without progression of neurologic exam. Fail-

ure of medical treatment was defined as worsening neurologic exam, sepsis

or death or progression of SEA on radiographic findings despite one week

of intravenous antibiotics.

Refer to onsite Annual Meeting presentations and postmeeting proceedings for po

reporting disclosures and FDA device/drug

METHODS: Patient demographics, radiographic appearance and the clin-

ical course and treatment methods utilized for all patients admitted with

SEA were collected. Both univariate and multivariate analysis were uti-

lized to identify independent variables that determined success or failure

of initial medical management.

RESULTS: Three hundred fifty-five patients with average age of 60 years

met our inclusion criteria. Diabetes, alcohol abuse and intravenous drug

use were the most statistically significant risk factors for developing

SEA. In-hospital mortality rate for the entire cohort of SEA was 8.5%;

mortality rate within 90 days from admission was 13.1%. One hundred

forty-two patients were initially included in the nonoperative cohort. How-

ever, after one week of antibiotics, 42 patients required surgical decom-

pression. Twelve patients died from failure of medical treatment. In

sum, 54 patients failed medical management and 73 patients were success-

fully treated without surgical intervention. Univariate analysis identified

age, neurologic status at the time of presentation, diabetes, epidural ab-

scess located above the level of the conus medullaris and circumferential

epidural abscess as significant risk factors leading to failure of medical

treatment. Multivariate analysis identified incomplete or complete spinal

cord deficits as the most significant risk factor for unsuccessful medical

management. Age greater than 65 years, diabetes and methicillin resistant

staphylococcus aureus (MRSA) were also independent risk factors for fail-

ure of medical management. An algorithm for probability of failed nonop-

erative management of spinal epidural abscess predicted 99% probability

of failure for patients with all four of these risk factors.

CONCLUSIONS: SEAs treated with medical management alone have

a very high risk for failure in patients with diabetes, MRSA infection, neu-

rologic compromise, and 65 years of age or older. In the absence of these

independent risk factors, nonoperative management of SEA may be con-

sidered as an initial treatment provided close clinical monitoring can be

delivered.

FDA DEVICE/DRUG STATUS: This abstract does not discuss or include

any applicable devices or drugs.

http://dx.doi.org/10.1016/j.spinee.2013.07.136

89. Risk Factors for a Delay in Diagnosis of Vertebral Osteomyelitis

Sina Pourtaheri, MD1, Mark J. Ruoff, MD2, Eiman Shafa, MD3,

Arash Emami, MD4, Tyler N. Stewart5, Kimona Issa, MD6, Ki S. Hwang,

MD7, Kumar G. Sinha, MD4; 1Teaneck, NJ, US; 2Orthopaedic Associates,

Fair Lawn, NJ, US; 3Saddle Brook, NJ, US; 4University Spine Center,

Wayne, NJ, US; 5New York, NY, US; 6Baltimore, MD, US; 7University

Place Spine Center, Wayne, NJ, US

BACKGROUND CONTEXT: Vertebral osteomyelitis, especially Pott’s

disease, has been well established as having a delay in diagnosis due to

the specific complains and findings on presentation.

PURPOSE: The purpose of this study was to identify clear risk factors for

a delay in diagnosis for vertebral osteomyelitis.

STUDY DESIGN/SETTING: Retrospective clinical and radiographic

review.

PATIENT SAMPLE: 920 patients from a single institution who had ver-

tebral osteomyelitis from 2001to 2011.

OUTCOME MEASURES: Clearance of the infection, length of delay of

diagnosis, Nurick grade, Oswestry score, segmental kyphosis, length of

hospital stay (LOS), cost of hospital admission, mortality.

METHODS: A retrospective review of 920 patients who had vertebral os-

teomyelitis was performed. Inclusion criteria included appropriate initial

imaging, lab results, evaluation by the orthopedic department, and no treat-

ment done prior to admission at an outside institution. Chi-squared statis-

tic, Fisher’s exact, and single sample t tests were used to examine the data.

All patient characteristics were evaluated as potential risk factors. A delay

of diagnosis was defined as greater than 8 weeks from first ER visit to di-

agnosis. Clearance of the infection was defined as normalizing of serum

markers and resolution of osteomyelitis on MRI after 6 months of

ssible referenced figures and tables. Authors are responsible for accurately

status at time of abstract submission.